False-positive human immunodeficiency virus enzyme immunoassay results in pregnant women

Objective

Examine whether false-positive HIV enzyme immunoassay (EIA) test results occur more frequently among pregnant women than among women who are not pregnant and men (others).

Design

To obtain a large number of pregnant women and others tested for HIV, we identified specimens tested at a national laboratory using Genetic Systems HIV-1/HIV-2 Plus O EIA from July 2007 to June 2008.

Methods

Specimens with EIA repeatedly reactive and Western blot-negative or indeterminate results were considered EIA false-positive. We compared the false-positive rate among uninfected pregnant women and others, adjusting for HIV prevalence. Among all reactive EIAs, we evaluated the proportion of false-positives, positive predictive value (PPV), and Western blot bands among indeterminates, by pregnancy status.

Results

HIV prevalence was 0.06% among 921,438 pregnant women and 1.34% among 1,103,961 others. The false-positive rate was lower for pregnant women than others (0.14% vs. 0.21%, odds ratio 0.65 [95% confidence interval 0.61, 0.70]). Pregnant women with reactive EIAs were more likely than others (p<0.01) to have Western blot-negative (52.9% vs. 9.8%) and indeterminate results (17.0% vs. 3.7%) and lower PPV (30% vs. 87%). The p24 band was detected more often among pregnant women (p<0.01).

Conclusions

False-positive HIV EIA results were rare and occurred less frequently among pregnant women than others. Pregnant women with reactive EIAs were more likely to have negative and indeterminate Western blot results due to lower HIV prevalence and higher p24 reactivity, respectively. Indeterminate results may complicate clinical management during pregnancy. Alternative methods are needed to rule out infection in persons with reactive EIAs from low prevalence populations.     

The kinetics of activation of normal and gamma-carboxyglutamic acid-deficient prothrombins

The kinetics of activation of normal and gamma-carboxyglutamic acid (Gla)-deficient prothrombins isolated from cattle maintained for extended periods on the vitamin K antagonist dicoumarol were studied. The catalyst was prothrombinase, comprising isolated Factor Xa, Factor Va, phospholipid vesicles, and calcium ion. The Km and kcat values for prothrombins with 0, 1, 2, 5, 7, and 10 Gla residues were determined both by initial rate analysis and by integrated Michaelis-Menten-Henri analysis. Each of the Gla-deficient prothrombins exhibited kcat values similar to that of normal 10-Gla prothrombin but Km values that were 8- to 20-fold greater than that of the normal molecule. The increased Km coincided with a loss of Ca2+- and phospholipid-binding properties of the Gla-deficient prothrombins. The magnitude of the defect in both the kinetics of activation and Ca2+ and phospholipid binding is not progressive with the loss of Gla residues but rather appears abruptly with the loss of as few as 3 of the 10 Gla residues present in the normal substrate. The theoretical relationship between Km(app) and the dissociation constant (Kd) of the prothrombin-phospholipid interactions was derived. According to the result, the increase in apparent Km observed with the Gla-deficient prothrombins corresponds to at least a 100- to 1000-fold decrease in affinity for phospholipid compared to the affinity of normal prothrombin. In addition, the products of the activation of 10-Gla prothrombin were found to inhibit the activation of the Gla-deficient prothrombins.     

Evidence that the thrombin-catalyzed feedback cleavage of fragment 1.2 at Arg154-Ser155 promotes the release of thrombin from the catalytic surface during the activation of bovine prothrombin

During the course of prothrombin activation, as catalyzed by Factor Xa, Factor Va, Ca2+, and negatively-charged phospholipid vesicles, the three proteins distribute between the fluid phase and the vesicle surface. On the vesicle, efficient Factor Xa-catalyzed proteolysis yields thrombin plus Fragment 1.2. Further thrombin-catalyzed feedback cleavage of the latter then yields Fragment 1 plus Fragment 2. Prior to this cleavage Fragment 1.2 might retain thrombin at the site of catalysis since it binds both phospholipid and thrombin through its respective Fragment 1 and Fragment 2 domains. In order to study the role of the feedback cleavage, light scattering at right angles was used to deduce the nature of the components associated with the vesicle during prothrombin activation by continuous monitoring of the relative molecular weight of the vesicle-protein complex. When prothrombin (1.4 microM) was added to homogeneously sized phospholipid vesicles of phosphatidylcholine-phosphatidylserine (3:1) at a total phospholipid concentration of 20 microM, the scattering intensity doubled. Upon subsequent addition of Factor Xa and Factor Va (5.0 nM each) the scattering intensity smoothly decreased to a value about 1.25-fold greater than that of the vesicles alone. Analysis of the composition of the reaction mixture at intervals during the course of the reaction by gel electrophoresis and laser densitometry, provided a good correlation between the mass of the vesicle-protein complex measured by light scattering and its mass inferred by composition. In addition, the decrease in mass of the vesicle-protein complex measured by light scattering correlated temporally with cleavage of Fragment 1.2. When the reaction was initiated in the presence of the reversible thrombin inhibitor dansylarginine-N-(3-ethyl-1,5-pentanediyl)amide no cleavage of Fragment 1.2 occurred, as indicated by gel electrophoresis, and no change in the mass of the vesicle-protein complex occurred as indicated by light scattering. The absence of change in scattering intensity in the presence of dansylarginine-N-(3-ethyl-1,5-pentanediyl)amide suggests a 1:1 replacement of prothrombin at the catalytic surface by components of equivalent mass (Fragment 1.2 plus thrombin), whereas the decrease in scattering in the absence of dansylarginine-N-(3-ethyl-1,5-pentanediyl)amide suggests replacement of prothrombin by Fragment 1 only. Together these results indicate that the thrombin-catalyzed cleavage of Fragment 1.2 promotes release of thrombin from the catalytic surface.     

Activation of human prothrombin by human prothrombinase. Influence of factor Va on the reaction mechanism

The kinetics of the activation of human prothrombin catalyzed by human prothrombinase was studied using the fluorescent alpha-thrombin inhibitor dansylarginine-N-(3-ethyl-1,5-pentanediyl)amide (DAPA). Prothrombinase proteolytically activates prothrombin to alpha-thrombin by cleavages at Arg273-Thr274 (bond A) and Arg322-Ile323 (bond B). The differential fluorescence properties of DAPA complexed with the intermediates and products of human prothrombin activation were exploited to study the kinetics of the individual bond cleavages in the zymogen. When the catalyst was composed of prothrombinase (human factor Xa, human factor Va, synthetic phospholipid vesicles, and calcium ion), initial velocity studies of alpha-thrombin formation indicated that the kinetic constants for the cleavage of bonds A or B were similar to the constants that were obtained for the overall reaction (bonds A + B). The progress of the reaction was also monitored by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. The results indicated that the activation of human prothrombin catalyzed by prothrombinase proceeded exclusively via the formation of meizothrombin (bond B-cleaved) as an intermediate. Kinetic studies of the cofactor dependence of the rates of cleavage of the individual bonds indicated that, in the absence of the cofactor, cleavage at bond B would constitute the rate-limiting step in prothrombin activation. Progress curves for prothrombin activation catalyzed by prothrombinase and monitored using the fluorophore DAPA were typified by the appearance of a transient maximum, indicating the formation of meizothrombin as an intermediate. When factor Xa alone was the catalyst, progress curves were characterized by an initial burst phase, suggesting the rapid production of prethrombin 2 (bond A-cleaved) followed by its slow conversion to alpha-thrombin. Gel electrophoresis followed by autoradiography was used to confirm these results. Collectively, the results indicate that the activation of human prothrombin via the formation of meizothrombin as an intermediate is a consequence of the association of the cofactor, human factor Va, with the enzyme, human factor Xa, on the phospholipid surface.     

Long-term effects of estrogen on avian liver: estrogen-inducible switch in expression of nuclear, hormone-binding proteins.

The stimulation of chicks or embryos with estrogen results in transient, hepatic expression of the vitellogenin gene, as well as long-term, propagatable alterations in the rapidity with which the gene can be reactivated. We examined the possibility that nuclear, type II estrogen-binding sites are involved in this long-term change in response characteristics. We demonstrate that the primary induction kinetics of type II sites in embryos and chicks correlated with the expression of the vitellogenin gene and that once their induction was triggered by estrogen, they accumulated, were propagated, and persisted for months after withdrawal of the hormone. We also show that their accumulation in the embryo was accompanied by prolonged expression of both the vitellogenin and very low-density apolipoprotein II genes, in the absence of elevated levels of type I receptor, and that the type II sites, like the classical receptor, appear to be preferentially associated with active or potentially active chromatin. Finally, we describe a regulatory mechanism, tested by computer modelling, that simulated the behavioral characteristics of these nuclear estrogen-binding sites and which may explain their role in mediating the long-term effects of estrogen.     

The prothrombinase-catalyzed activation of prothrombin proceeds through the intermediate meizothrombin in an ordered, sequential reaction

The activation of bovine prothrombin by prothrombinase (Factor Xa, Factor Va, synthetic phospholipid vesicles, and calcium ion) was studied in the presence of the fluorescent, reversible thrombin inhibitor dansylarginine-N-(3-ethyl-1,5-pentanediyl) amide (DAPA). Recordings of fluorescence intensity during prothrombin activation exhibited maxima that decreased to stable limiting values. These data suggested the transient appearance of the meizothrombin-DAPA complex, which exhibits fluorescence with 1.5-fold greater intensity than the thrombin-DAPA complex. At substrate concentrations well below Km, progress curves could be fitted by equations describing an ordered, sequential conversion of prothrombin to thrombin through the intermediate meizothrombin via two pseudo-first order steps. The pseudo-first order rate constants for both steps varied linearly with enzyme concentration, indicating that both steps are catalyzed by prothrombinase. The progress of the reaction was also monitored by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and densitometry analyses of aliquots removed at intervals spanning the reaction. These analyses confirmed both the existence of meizothrombin and its time course as predicted from the equations used to analyze fluorescence intensity profiles. Meizothrombin levels peaked at about 0.3 mol/mol initial prothrombin under the conditions typically studied. In addition, prethrombin 2, which is the intermediate expected from cleavages occurring in the order opposite that required to form meizothrombin, was not observed under any of the conditions examined. These data indicate that prothrombin activation catalyzed by the fully assembled prothrombinase complex proceeds via an ordered, sequential reaction with meizothrombin as the sole intermediate.     

Differentiation of enzyme and substrate binding in the prothrombinase complex

The Ca2+ dependence of factor Xa binding to phospholipid vesicles was measured in the presence and absence of factor Va. The increase in polarization of a fluorescently labeled derivative of factor Xa, [5-(dimethylamino)-1-naphthalenesulfonyl] glutamylglycylarginyl factor Xa (Dns-EGR-Xa), was used as a probe to measure the interaction of factor Xa with phospholipid. The Ca2+ concentration required for half-maximal binding of Dns-EGR-Xa to phospholipid vesicles was 3.5 X 10(-4) M in the presence of factor Va and 9.5 X 10(-4) M in the absence of factor Va. At a Ca2+ concentration of 5 X 10(-4) M, the binding of Dns-EGR-Xa to phospholipid-bound factor Va was near maximal, whereas there was no detectable interaction of Dns-EGR-Xa with phospholipid alone at this Ca2+ concentration as detected by fluorescence polarization. These results were qualitatively confirmed by high-performance liquid chromatography. The rate of hydrolysis of the factor Xa synthetic substrate, benzoylisoleucylglutamylglycylarginine p-nitroanilide, by factor Xa in the presence of factor Va and phospholipid decreased in a Ca2+-dependent manner. These data were analyzed as fraction of factor Xa bound to the phospholipid. A Ca2+ concentration of 2.7 X 10(-4) M resulted in half-maximal binding by this technique. The relationship observed between rates of prothrombin activation and Ca2+ concentration could be predicted quantitatively from calculations of local enzyme and substrate concentrations.